AI

AI Economics Featured Military The International Economy

The Bubble That Could Shape the Century

The AI-Bubble

“Artificial Intelligence (AI) is likely to be either the best or the worst thing to happen to humanity,” says entrepreneur and trillionaire Elon Musk. Musk, the CEO of SpaceX and Tesla, is known for his bold and often controversial statements on the future of technology. He has been a vocal advocate for developing AI, but has also warned about the potential dangers it could pose to humanity if not properly managed. While most discussions focus on AI’s direct social impacts, like automation and job insecurity, the bigger risks lie in its economic and institutional effects on governments and markets.

Recently, massive investments in AI have fueled what many call an “AI bubble,” driven by projected future returns, inflated stock prices, and a gap between the hype and the actual, sustainable profitability of AI technologies and business models. Experts are divided. Some see a bubble forming, while others argue that AI’s tangible utility and the financial strength of tech giants make the boom sustainable. Recent market movements underscore this tension. Investor Michael Burry, famous for predicting the 2008 financial crisis, has taken high-profile short positions against AI giants like Nvidia and Palantir, even as these companies continue to post record-breaking revenues and attract billion-dollar investments from Microsoft, Google, Meta, OpenAI, and major venture capital firms.

Between these companies, there is a growing competition for human capital. Major corporations have begun aggressively poaching AI researchers and engineers from one another, highlighting both the scarcity and the value of expertise. For example, Apple has recruited numerous AI specialists from Google, establishing a new research hub in Zurich, while Meta’s “superintelligence” division has drawn top talent from DeepMind and Scale AI. These moves are backed by staggering compensation packages. Meta’s top AI engineers earn base salaries up to $440,000, while senior researchers at OpenAI can make between $800,000 and $1.3 million annually when including equity and bonuses. Apple’s AI engineers, too, can reach base salaries above $300,000, reflecting a market where talent itself has become an asset class. This scramble for expertise not only concentrates innovation within a handful of dominant firms but also raises barriers for smaller competitors and public institutions.

These patterns mirror the 1990s Dot-com bubble, where companies with no profits received enormous investments and evaluations surrounding the internet’s potential, which resulted in a market collapse in 2000. While that crash resulted in significant financial losses, job cuts, and the failure of many dot-com companies, it nevertheless led to a more cautious approach in the tech industry, with a greater emphasis on profitability and stricter regulations on financial reporting and disclosure to protect investors. Critics of the AI bubble turn to these provisions to defend the investment in the AI market. They argue that AI is already delivering real value across various industries, where investment is coming from large, profitable tech companies with strong balance sheets. The AI industry’s growth is projected to continue for years, suggesting that the boom may not be a bubble but rather a genuine technological revolution. Yet, even if the boom isn’t a bubble, AI-driven investment is reshaping global markets and widening the world’s militant and economic divide.

AI’s Effect on Global Markets

AI-driven investment is heavily concentrated in a few countries, creating profound effects on international markets and economic power structures. In 2025, the United States leads global AI spending with approximately $470 billion, followed by China at $119 billion, the United Kingdom at $28 billion, and countries like Canada, Israel, Germany, and India trailing behind. This concentration gives these countries outsized influence over global trade in AI-related goods and services, including cloud computing, semiconductors, AI software, and high-performance computing resources. Companies and governments in these leading nations are able to set standards, control supply chains, and dictate pricing, while countries with limited investment struggle to compete in these emerging markets.

The imbalance also affects international capital flows. Wealthier nations attract the bulk of AI venture capital and access to AI infrastructure, where expertise remains concentrated in a few economies. Currently, just “100 firms, mainly in the US and China, account for 40% of global corporate research and development (R&D) spending,” accumulating a “market value of around $3 trillion, rivalling the gross domestic product of the whole African continent.” The U.S.’s partnership with Nvidia, which controls 85-90% of the global market for AI chips, gives the U.S. singular leverage over compute access. Similarly, China’s “Digital Silk Road” exports AI technology to more than 80 countries across Africa, Latin America, and Eastern Europe. India and African economies rely on imported computing capacity, where the majority of the networks for AI infrastructure come from companies headquartered in the U.S. or China.

This creates a feedback loop: countries that already dominate global markets receive more investment, strengthen their technological infrastructure, and expand influence, while smaller economies remain technologically dependent and economically marginalized. The result is a reinforced global hierarchy, where AI investment is not only a domestic economic strategy but also a tool for asserting international market dominance and shaping the rules of trade, intellectual property, and technology access.

AI’s Effects on Military Development

Beyond investment, AI’s rapid development has also transformed the global security landscape, intensifying competition between major powers and redefining the nature of warfare. Wealthier nations are investing heavily in AI-enabled defense systems, including autonomous drones, predictive intelligence platforms, cyberwarfare tools, and advanced command-and-control networks. The U.S., through initiatives like the Joint Artificial Intelligence Center (JAIC) and partnerships with OpenAI, integrates AI into both battlefield and intelligence operations, while China’s People’s Liberation Army (PLA) pursues “intelligentized warfare” to fuse AI with military command systems. For example, the U.S. uses programs like Project Maven for real-time target identification; DARPA tests autonomous drone swarms; and the Pentagon increasingly relies on AI-enabled command-and-control systems. Other nations, including Russia, Israel, and the U.K., are also developing AI-driven defense capabilities, but their investments are far smaller, leaving them dependent on alliances or technology imports. Israel’s IDF employs AI systems such as “Gospel” and “Fire Factory” for automated target selection and operational planning and exports AI-enabled drones to India, Azerbaijan, and the UAE, extending its influence. Meanwhile, Russia is experimenting with autonomous ground vehicles (Uran-9) in Syria and AI-assisted cyberwarfare tools used in operations against Ukraine.

This concentration of AI military capacity has significant diplomatic implications. Countries controlling advanced AI capabilities can leverage them to exert influence in international negotiations, peacekeeping, and strategic alliances. For example, a nation with superior autonomous surveillance and cyber capabilities may have greater leverage in trade agreements or conflict mediation. Smaller states or developing countries, especially in Africa, the Middle East, and Southeast Asia, are lacking domestic AI resources, are often forced into strategic dependence, relying on AI technologies from wealthier powers and limiting their ability to act independently in international affairs. Analysts warn that this uneven distribution risks creating a new global hierarchy of influence, where technological superiority translates directly into diplomatic and geopolitical advantage, reinforcing existing inequalities between the Global North and South.

Mitigating Global AI Inequality

These economic and military transformations ultimately converge into a broader global inequality problem where the development of AI is concentrated in a small group of wealthy states. The United States and China dominate AI infrastructure, research, and talent, giving them disproportionate control over markets, trade flows, and international standards. This accelerates their economic growth while limiting opportunities for developing countries, creating a technological and economic divide, and fostering dependency on imported AI systems. The same inequality extends to military and diplomatic power; nations with advanced AI defense systems, including autonomous weapons, cyberwarfare tools, and predictive intelligence platforms, can project influence globally, shaping negotiations, alliances, and conflict outcomes. Smaller or developing countries face strategic dependence, reducing their autonomy in international affairs and reinforcing global hierarchies (Europarl, 2024; Al Jazeera, 2025).

Yet, this trajectory is not inevitable. The AI bubble highlights both the risks of overconcentration and the potential for disproportionate influence by a few nations and firms. Global coordination, equitable investment in infrastructure, open-access AI initiatives, and international standards for ethical deployment can help mitigate disparities. Programs such as the EU Digital Europe Initiative, Horizon Europe, and UN-supported collaborative research provide models for sharing expertise, funding, and technology. Lessons from previous technological revolutions, like the internet and mobile communications, suggest that deliberate policy and international cooperation can prevent concentrated power from solidifying into long-term inequality. If nations act strategically, it is possible to balance innovation with equity, ensuring that AI delivers broad economic, social, and security benefits rather than reinforcing existing hierarchies.

Image source: public domain pictures

November 25, 2025 0 comment

AI Climate Change Environmental Change South America

Impacts of ChatGPT on the Lithium Triangle as Mining Efforts Scale Up

written by Georgiana Morris

One of the largest breakthroughs of the 21st century has been in the field of artificial intelligence (AI), particularly the creation of ChatGPT. Widely known for its intuitive language processing and free, public access to an unlimited amount of information, ChatGPT has become ingrained into daily life, and as it continues to evolve and expand its influence, we must be aware of not just what it offers us but what it also demands from others. Injustices caused indirectly by ChatGPT have long harmed local indigenous and rural communities, meaning that AI’s explosive growth threatens to worsen already precarious conditions. To recognize the impacts of ChatGPT on these communities, there are two essential components to understand first: the resource-intensive nature of both ChatGPT and lithium mining.

Revolutionary in its making, Large Language Models (LLM) like ChatGPT undergo a series of machine learning training stages to allow consumers access to a seemingly unlimited supply of information. During its unsupervised training phase, where the machine is given access to a large corpus of text, LLMs learn general features and patterns of natural language. This method is known as a transformer-based neural network and includes processing web sources, books, online datasets, and human-generated data. Training these models demands huge amounts of energy, as data centers rely heavily on water and cooling systems, causing significant environmental damage. The Washington Post found that for every 100-word email generated by an AI chatbot, about 1 bottle of water is wasted. Think of the water bill for that!

Now that we understand ChatGPT and its cost of running, let’s look at the cost of creation: lithium extraction. Lithium is one of many “critical minerals”, metals essential for battery technologies like electric vehicles (EVs), renewable energies (REs), and data centers. The global supply chain of lithium is vast, interconnected, and vulnerable. Factors like geographic concentration of extraction and the associated “arms race” to control these areas mean that local communities are often overlooked and exploited. This trend was exacerbated by Donald Trump in 2017, when he signed an executive order expanding the concept of critical mineral security to include the entire supply chain, in an effort to assert US influence and dominance in controlling global resources. This had stark consequences, as the influx of countries competing to control these resources further deteriorated any regard for localized impact.

In one such location, the Lithium Triangle, injustices for indigenous communities are extremely evident, particularly around the mining sites. Nestled between the borders of Argentina, Bolivia, and Chile, this area of the Andes hosts more than 75% of the world’s lithium reserves and is essential to the creation and maintenance of AI. However, many of these deposits are located on the ancestral lands of indigenous communities, who have consistently faced threatened ecosystems from lithium mining. Raw lithium is primarily extracted by drilling into subsurface fluids, known as brines, and pumping the liquid into pools to evaporate the water, leaving behind the lithium. This process is highly energy-intensive and causes immense socio-environmental costs, such as water depletion, ecosystem destruction, indigenous displacement, and economic inequity. These costly tendencies are labeled green extractivism, a term reminiscent of past extractivist colonial institutions, as corporations override sustainability concerns through lobbying. Often, local governments ease indigenous land rights and restrict the ability to protest in favor of profits. Historically, the indigenous communities hosting the mining industries have been disproportionately impacted, known as green sacrificial zones (GSZ), areas prone to chemical pollution from industries. Lithium mining produces these GSZs, where private enterprises shift the economic and environmental costs of production to neighboring communities, known as frontline communities, who suffer directly. For example, in the Puna Plateau in Argentina, the indigenous frontline community relies on underground water adjacent to the salt flats, which is continuously depleted by mining efforts in that region. Extractivist tendencies, as observed from the European conquest of Africa in the 1800s, have preyed on and burdened these communities with impossible impacts, without installing proper infrastructure to involve community leaders in mining processes.

The ever-increasing rise of language models has accelerated the demand for lithium batteries to power data centers. This means that global demand for lithium has increased, scaling up mining efforts despite the vulnerable communities immediately affected. The just transition is built on fossil fuels, as all energy transitions before it built upon their energy predecessors. So even while lithium is used in EVs and other RE strategies, its cost is rooted in inequality. Therefore, as OpenAI products are being scaled up to meet the high demand for advancements in the field, and more data centers are built to supply the demand, more lithium-ion batteries are created within the already vulnerable global market, further damaging the frontline communities of the mining industry.

Lithium mining has serious environmental and economic issues already prevalent in many green sacrificial zones abutting mining areas. Ratifying human rights principles like Free, Prior and Informed Consent (FPIC) helps enshrine accountability from tech companies as they involve local indigenous communities in mining processes and share in economic benefits. Stricter environmental and labor regulations in lithium mining are integral to indigenous community rights protection as well as investment in alternative battery chemistries and locally-led RE initiatives or mining cooperatives. We must ask ourselves: Are we willing to pay the cost of our queries, or let others bear the burden in their quarries?

May 18, 2025 302 comments

AI Nuclear Deterrence Technology

Pumping the Brakes on Artificial Intelligence in Nuclear Command and Control

written by Charlie Donlan

In a meeting on November 16, 2024, U.S. President Joe Biden and Chinese President Xi Jinping found common ground on artificial intelligence (AI), recognizing the need for human beings to have ultimate authority over decisions to use nuclear weapons. Their agreement represents a rare moment of cooperation between two countries that are constantly competing in both AI and nuclear innovation. The rapid advancement in the capabilities of artificial intelligence and its increasing accessibility to the public is becoming a defining theme of the 21st century. The development of hyper-advanced machine learning algorithms, like those used in generative AI models like GPT 4, continues to increase the autonomy of artificial intelligence. As computers become less reliant on human intelligence, humans become more reliant on computers. Meanwhile, advancements in nuclear technology and renewed nuclear proliferation now tempt state actors to leverage the machine learning power of artificial intelligence and automate the decision to use a nuclear weapon. Given these incentives and the current capabilities of AI, the decision world leaders are faced with is not if AI should be used in nuclear command and control but to what extent.

In order to properly assess the risks of relying on the decision-making of artificial intelligence when it comes to nuclear weapons use, an understanding of existing nuclear weapons systems and the factors that would incentivize states to automate nuclear command and control protocols is necessary. The U.S. Nuclear Command, Control, and Communications (NC3) system serves five main functions for the U.S. nuclear weapons program:

detection, warning, and attack characterization
nuclear planning
decisionmaking conferencing
reception of presidential orders
management and direction of forces

These primary functions are executed according to a rigid chain of command system involving a number of different government and military agencies. Ultimately, the President has the sole authority to authorize the use of a nuclear weapon.

Since the Cold War, the U.S. and other nuclear weapons states have used artificial intelligence to improve efficiency in various sectors of the NC3 system. For example, the U.S. uses artificial intelligence automation in its early-warning systems to detect an incoming nuclear attack and guide outgoing attack missiles. These types of artificial intelligence tools generally belong to a category known as “rule-based AI,” which refers to AI models that operate based on direct programming from a human source. Currently, rule-based AI in nuclear weapons procedures is used in conjunction with human decision-making.

Artificial intelligence applications in nuclear command and control are contingent upon the principle that nuclear weapons states would only detonate a nuclear weapon in response to a foreign nuclear attack. Currently, the only nuclear weapons state with a comprehensive “no first use” policy is China. India also has a conditional “no first use” policy that warrants a nuclear response to chemical or biological weapons attacks. Despite attempts to pass legislation declaring a “no first use” policy in the U.S., current law allows for the U.S. to preemptively strike an adversary with nuclear weapons. However, former Defense Secretary William Perry acknowledges that no “rational president” would ever authorize the first use of nuclear weapons.

Because nuclear weapons states have demonstrated an adherence to the doctrine of “no first use”, the need for timely response to incoming nuclear attacks is critical. With the current state of nuclear weapons technology, governments would likely have only minutes to detect and respond to nuclear threats before they are attacked. However, the decision on how to proceed once an attack has been detected has become more and more difficult due to the proliferation of nuclear weapons. During the Cold War, the U.S. military was able to develop detailed plans for responses to nuclear attacks because their only threat was the Soviet Union. Today, there are nine countries that possess viable nuclear weapons. Additionally, intelligence on adversarial nuclear powers has also become more ambiguous. For example, North Korea has been known to conceal its nuclear weapons program from international inspection and intelligence. Growing uncertainty regarding nuclear threats and the need for timely response to an attack has forced nuclear weapons states to look for ways to increase the efficiency of their nuclear response capabilities.

Existing difficulties in nuclear command and control operations cast additional doubts on the potential for safe integration of artificial intelligence. In 2015, The Union of Concerned Scientists published a report detailing historical examples of errors in command and control that nearly resulted in incidental or ill-advised nuclear detonation.

The first category of errors identified in the report related to ballistic missiles are instances of ambiguous data generated by early warning sensors. An example of this type of error was in 1995 when Russian early warning systems detected a Norwegian scientific rocket and classified it as a nuclear threat, nearly leading to a retaliatory strike from Russia. Even though Russia was previously notified about the launch, there was a lapse in communication.

Another category of errors identified in the report is false alarms caused by natural phenomena. In 1983, a Soviet early warning satellite detected five land-based missiles launched by the U.S. at the Soviet Union, which turned out to be reflections of the sun on the clouds that misled the missile detection system. According to the report:

“Since the satellite was found to be operating properly, following procedures would have led him [the Soviet officer on duty] to report an incoming attack. Going partly on gut instinct and believing the U.S. was unlikely to fire only five missiles, he told his commanders that it was a false alarm before he knew that to be true.”

In this case, a nuclear strike was saved by the “gut instinct” of the Soviet officer. Clearly, there are many challenges to the nuclear command and control framework exemplified by both human error and machine error.

While nuclear weapons states demonstrate a growing interest in leveraging the power of AI to improve the efficiency and accuracy of nuclear response, AI developers are showing more of an interest in military applications for their products. As mentioned, basic forms of artificial intelligence such as Rule-based AI have been used in nuclear command and control procedures since the Cold War era. Rule-based AI mimics human decision-making by operating based on conditions directly programmed by humans. This kind of artificial intelligence uses a network of syllogistic “if-then” instructions to make decisions based on input data. Because of the simplicity of the programmed instructions and the clarity in the decision-making process of the model, rule-based AI typically poses minimal risk of error. However, as the landscape of nuclear threats widens, leaders are looking at more advanced AI models that are able to process larger amounts of data more efficiently.

Recent breakthroughs in artificial intelligence have been centered around advancements in machine learning. Machine learning models differ from rule-based models in the source of their programming. Machine learning can be generally defined as “the capability of a machine to exhibit intelligent human behavior” – integral to that is the model’s ability to teach itself in the absence of human programming. Instead of operating based on a stagnant knowledge base of instructions, a machine learning model is able to constantly update its own instructions based on new data it receives.

There is no doubt that recent advancements in machine learning have made the world a safer place in many respects. Thanks to modern machine learning, self-driving cars can keep passengers safer than any human driver possibly could, banks can more effectively detect credit card fraud, and medical imaging can be more accurately analyzed to diagnose illnesses. Despite the many positive impacts modern machine learning AI has had on human lives, there are still several technical shortcomings.

The first technical issue that is characteristic of modern AI models is human bias. While the algorithms used by modern machine learning models have innate capabilities to self-modify, they are originally programmed by human beings. Therefore, the algorithms themselves can be subject to the same bias as the humans who created them. While this is also true of rule-based AI models, machine learning algorithms have the ability to extrapolate human bias in their programming. In generative AI models that use machine learning algorithms, such as GPT 4, human bias could be applied in unintended situations that prove dangerous. There are countless examples of human bias impacting AI outputs, such as the model Amazon used to filter resumes, which had a predetermined bias toward male applicants.

A 2022 report by Arms Control Today outlines the risks of human bias on AI as it relates to nuclear weapons. Modern machine learning algorithms are not developed by people with knowledge about nuclear threat detection, meaning the AI developer’s lack of understanding about nuclear command and control might create a human bias in their model. Although an AI model might proceed flawlessly according to its instructions, bias in the instructions themselves could lead to unpredictable outputs.

Another problem with modern AI models pertains to outputs generated from inaccurate information. A concerning characteristic of modern AI models is that their decision-making is solely based on the data they receive. This means that there is no way for an AI program to determine whether the data it receives is accurate or best suited for the job it is designed to perform. Imagine an AI model designed to read radar data and detect incoming missile attacks. If the model has been trained, even if inadvertently, using data from inaccurate radar readings, then it may produce an output based on the faulty data.

Even if an AI model is trained with accurate and relevant data, its algorithm still has the ability to produce unpredictable outputs that are unsupported by the training data. Applied to incoming missile detection, an AI model could be trained with completely accurate radar data and still produce an unintended output that may falsely identify or fail to detect a threat. AI outputs that do not align with the model’s provided training data are known as “hallucinations.”

The increasing complexity of modern machine learning algorithms come with increasing ambiguity over the steps AI models take to produce an output. The clarity of an AI model’s decision-making process is referred to by experts as “explainability.” In a highly explainable AI model, it is easy to discern each step the model takes from receiving data to generating a response. Models that use multi-layered machine learning algorithms are less transparent than those that use very simplistic algorithms, like rule-based AI models. Making decisions based on unexplainable AI-generated outputs would strip humans of the ability to check for errors in reasoning and refine models for future use.

Finally, unlike human intelligence, AI is vulnerable to external cybersecurity threats and attacks. The National Institute of Standards and Technology issued a report in January outlining the various methods and means used by adversaries to disrupt or influence complex machine learning models. The report breaks down cybersecurity attacks targeting AI models into four categories: evasion, poisoning, privacy, and abuse. These categories of cyber attacks involve adversaries gaining access to or manipulating AI training materials and input data with the goal of finding weaknesses in the model or generating unintended outputs. These cybersecurity threats call into question the extent to which AI should be implemented in nuclear command and control systems.

Artificial intelligence is very good at presenting data to human beings in a way that is manageable. However, when it comes to decision-making, AI is flawed. Artificial intelligence applications in nuclear command and control networks should be focused on assisting in areas where humans have demonstrated errors in decision-making. Reliable and explainable artificial intelligence models should be used in conjunction with human analysts for nuclear detection, warning, and attack characterization. However, the ultimate decision to deploy or detonate a nuclear weapon should only ever be made by human beings.

February 24, 2025 325 comments